How sifat192/aircraft-damage-classification works
This repository is a notebook-based aircraft-damage image classifier using VGG16 transfer learning. Its documented workflow takes images through preprocessing and augmentation, trains a custom classification head over pretrained features, then evaluates and visualizes predictions. The README also documents prediction on unseen images. The notebook itself was not excerpted, so component paths point to the notebook and relationships below are supported by the README rather than inspected implementation details.
Main entry points: Aircraft_Damage_Classification_VGG16.ipynb Model Training
Components 11
Data Preparation
- Aircraft Images
- Image Preprocessing
Aircraft_Damage_Classification_VGG16.ipynb - Data Augmentation
Aircraft_Damage_Classification_VGG16.ipynb
Model Training
- VGG16 Features
Aircraft_Damage_Classification_VGG16.ipynb - Custom Classifier
Aircraft_Damage_Classification_VGG16.ipynb - Model Training
Aircraft_Damage_Classification_VGG16.ipynb
Evaluation and Prediction
- Model Evaluation
Aircraft_Damage_Classification_VGG16.ipynb - Accuracy and Loss
Aircraft_Damage_Classification_VGG16.ipynb - Unseen Image Prediction
Aircraft_Damage_Classification_VGG16.ipynb - Prediction Visualization
Aircraft_Damage_Classification_VGG16.ipynb
Other components
- User
Connections 11
- User → Aircraft Images: provides
README.md - Aircraft Images → Image Preprocessing: loads and processes
README.md - Image Preprocessing → Data Augmentation: augments training data
README.md - Data Augmentation → VGG16 Features: supplies images
README.md - VGG16 Features → Custom Classifier: extracts features for
README.md - Custom Classifier → Model Training: trains during
README.md - Model Training → Model Evaluation: evaluates
README.md - Model Evaluation → Accuracy and Loss: reports
README.md - Model Training → Unseen Image Prediction: enables
README.md - Unseen Image Prediction → Prediction Visualization: visualizes
README.md - User → Unseen Image Prediction: requests
README.md